On 1/19/21 12:30 PM, Xueming(Steven) Li wrote:

Hi Andrew,

quoted

-----Original Message-----
From: Andrew Rybchenko <redacted>
Sent: Tuesday, January 19, 2021 4:01 PM
To: Xueming(Steven) Li <redacted>
Cc: dev@dpdk.org; Slava Ovsiienko <redacted>; Asaf Penso
[off-list ref]; NBU-Contact-Thomas Monjalon
[off-list ref]; Ferruh Yigit [off-list ref]
Subject: Re: [PATCH v5 5/9] ethdev: support PF index in representor

On 1/19/21 10:14 AM, Xueming Li wrote:

quoted

With Kernel bonding, multiple underlying PFs are bonded, VFs come from
different PF, need to identify representor of VFs unambiguously by
adding PF index.

This patch introduces optional 'pf' section to representor devargs
syntax, examples:
 representor=pf0vf0             - single VF representor
 representor=pf[0-1]sf[0-1023]  - SF representors from 2 PFs

Don't we need
representor=pf3
i.e. without VF or sub-function?

Standalone PF not used by Mellnaox PMD, but should be supported.
Will update.

quoted

quoted

Signed-off-by: Xueming Li <redacted>
Acked-by: Viacheslav Ovsiienko <redacted>
Acked-by: Thomas Monjalon <redacted>
---
 doc/guides/prog_guide/poll_mode_drv.rst |  2 ++
 lib/librte_ethdev/ethdev_private.c      | 13 +++++++++++--
 2 files changed, 13 insertions(+), 2 deletions(-)

diff --git a/doc/guides/prog_guide/poll_mode_drv.rst

b/doc/guides/prog_guide/poll_mode_drv.rst
index 86e5867f1b..b2147aad30 100644

--- a/doc/guides/prog_guide/poll_mode_drv.rst
+++ b/doc/guides/prog_guide/poll_mode_drv.rst

@@ -382,6 +382,8 @@ parameters to those ports.
    -a DBDF,representor=sf[1,3,5]
    -a DBDF,representor=sf[0-1023]
    -a DBDF,representor=sf[0,2-4,7,9-11]
+   -a DBDF,representor=pf1vf0
+   -a DBDF,representor=pf[0-1]sf[0-127]

 Note: PMDs are not required to support the standard device arguments

and users  should consult the relevant PMD documentation to see support

devargs.

quoted

diff --git a/lib/librte_ethdev/ethdev_private.c

b/lib/librte_ethdev/ethdev_private.c
index d513f035d0..b9fdbd0f72 100644

--- a/lib/librte_ethdev/ethdev_private.c
+++ b/lib/librte_ethdev/ethdev_private.c

@@ -120,8 +120,8 @@ rte_eth_devargs_process_list(char *str, uint16_t

*list, uint16_t *len_list,

quoted

  *
  * Representor format:
  *   #: range or single number of VF representor - legacy
- *   vf#: VF port representor/s
- *   sf#: SF port representor/s
+ *   [pf#]vf#: VF port representor/s
+ *   [pf#]sf#: SF port representor/s
  *
  * Examples of #:
  *  2               - single

@@ -133,6 +133,14 @@ rte_eth_devargs_parse_representor_ports(char

*str, void *data)  {
 	struct rte_eth_devargs *eth_da = data;

+	if (str[0] == 'p' && str[1] == 'f') {
+		eth_da->type = RTE_ETH_REPRESENTOR_PF;
+		str += 2;
+		str = rte_eth_devargs_process_list(str, eth_da->ports,
+				&eth_da->nb_ports, RTE_MAX_ETHPORTS);

May be RTE_MAX_ETHPORTS -> RTE_DIM(eth_da->ports) ?

Same here, the dim could be different than MAX value.

Hold on, just for my understanding. The maximum says how
many entries could be added to the array. So, why?

quoted

quoted

+		if (str == NULL)
+			goto err;

Below we should not allow legacy VF syntax without "vf" prefix.

For backward compatibility, default numbers to "vf", otherwise some existing
app like OVS that working with VF will break.

I mean if new syntax is used (i.e. we have pfX prefix), we must
deny legacy syntax for VFs below.

quoted

quoted

+	}
 	if (str[0] == 'v' && str[1] == 'f') {
 		eth_da->type = RTE_ETH_REPRESENTOR_VF;
 		str += 2;

@@ -144,6 +152,7 @@ rte_eth_devargs_parse_representor_ports(char *str,

void *data)

quoted

 	}
 	str = rte_eth_devargs_process_list(str, eth_da->representor_ports,
 		&eth_da->nb_representor_ports, RTE_MAX_ETHPORTS);
+err:
 	if (str == NULL)
 		RTE_LOG(ERR, EAL, "wrong representor format: %s\n", str);
 	return str == NULL ? -1 : 0;